AI Voices
AI text-to-speech generator
Studio-grade AI voiceovers at roughly half of ElevenLabs' price
Reviewed by The Desk · Last verified July 2026
“A genuinely strong value play: Asyncflow gives you 1000+ voices, near-instant cloning and a streaming developer API at roughly half of ElevenLabs' rates. Quality is very good for the money, even if naturalness and language depth still trail the category leader.”
Skip it ifYou need the absolute best naturalness, fine-grained emotional control, or the widest language coverage for premium production — ElevenLabs still leads there — or you want a mature, standalone voice API with a long production track record.
What it is
AI Voices is the text-to-speech engine from Async (formerly Podcastle), built on the company's proprietary Asyncflow model. It converts written text or uploaded documents into studio-quality narration using a library of more than 1,000 voices spanning a range of tones and accents, with support for 15+ languages and pronunciation tested across 100+ languages and dialects. The product first shipped as Asyncflow v1.0 in March 2025 with 450+ voices and has since expanded.
Beyond raw generation, it leans on Async's editing pedigree: chat-based controls for fixing pronunciation and inserting pauses, a custom dictionary for technical terms, multi-character assignment for dialogue and audiobooks, and voice cloning that now needs only a few seconds of sample audio. A developer Voice API adds streaming output and integrations with LiveKit, Twilio, n8n and Pipecat, positioning it for voice agents and IVR as well as content creation.
The core pitch is value. Async prices speech at roughly $40 per 500 minutes against ElevenLabs' ~$99, and API usage starts around $0.50 per hour, with a small free tier to test. The trade-off is that naturalness, emotional nuance and language breadth still trail the frontier leaders, and the standalone voice tiers are folded into Async's broader creator platform, which makes the exact allowances harder to pin down.
Key features
- Asyncflow v1.0 proprietary TTS model
- 1000+ AI voices across tones and accents
- 15+ languages (100+ tested for pronunciation)
- Voice cloning from a few seconds of audio
- Chat-based pronunciation and pause editing
- Custom pronunciation dictionary for technical terms
- Multi-character voice assignment for dialogue
- PDF and document upload to audio
- Voice API with streaming support
Who it’s for
Narrating podcasts and audiobooks
Voiceovers for YouTube and marketing videos
E-learning and training narration
Accessibility and read-aloud content
Pros & cons
The good
- Markedly cheaper than ElevenLabs (~$40 per 500 minutes vs $99) with a real free tier to test
- 1000+ voice library plus voice cloning from just a few seconds of recorded audio
- Developer Voice API with streaming and LiveKit, Twilio, n8n and Pipecat integrations from ~$0.50/hour
The catch
- Naturalness, emotional range and language depth still trail ElevenLabs and other frontier TTS models
- Voice pricing is bundled into the broader Async platform, so standalone tiers and character caps are hard to parse
- Younger model and API with a shorter production track record than established voice vendors
Pricing
| Tier | Price | What you get |
|---|---|---|
| Free | Free | ~15 min (12k characters) lifetime TTS · Access to voice library |
| Essentials | $11.99/mo | ~2 hours (120k characters) per month · Voice cloning |
| Pro | $23.99/mo | ~10 hours (500k characters) per month · Advanced editing suite |
| Teams | $39.99/mo | ~40 hours (2M characters) per month · Collaboration for 2-5 users |
| Business | Custom | Custom volume · Advanced security and support |
Free tier — ~15 minutes (12k characters) lifetime
Alternatives to AI Voices
More AI Audio & Voice →
Questions people ask
Is AI Voices the same as Podcastle?
Yes. AI Voices is the text-to-speech product from Podcastle, which rebranded as Async — the podcastle.ai/ai-voices link now redirects to async.com. It is powered by the company's proprietary Asyncflow model.
How does the pricing compare to ElevenLabs?
It undercuts ElevenLabs significantly. Async charges roughly $40 per 500 minutes of speech versus about $99 for the same volume on ElevenLabs, and there is a small free tier of about 15 minutes to try it.
Does it support voice cloning?
Yes. Voice cloning is available on paid plans and only needs a few seconds of recorded audio, down from the roughly 70 sentences the earlier version required.
Is there an API for developers?
Yes. Async offers a Voice API with streaming and voice cloning, plus integrations with Pipecat, n8n, LiveKit and Twilio, with usage-based pricing starting around $0.50 per hour.